Progenitor Cell Mobilisation Peripheral Blood Progenitor Cell Mobilisation Alters Myeloid, but Not Erythroid, Progenitor Cell Self-Renewal Kinetics

Home , Progenitor cell

Bone Marrow Transplantation (2001) 27, 241–248  2001 Nature Publishing Group All rights reserved 0268–3369/01 $15.00 www.nature.com/bmt Progenitor cell mobilisation Peripheral blood progenitor cell mobilisation alters myeloid, but not erythroid, progenitor cell self-renewal kinetics

SB Marley1, JL Lewis1, B Zheng1, RJ Davidson1, JG Davis2, C McDonald2, FQB Alenzi1, JM Goldman1 and MY Gordon1

1Leukaemia Research Fund Centre for Adult Leukaemia and 2Stem cell Laboratory, Department of Hematology, Imperial College School of Medicine, Hammersmith Campus, London, UK

Summary: to cytokines may modify PBPC progenitor cell kinetics. Bone Marrow Transplantation (2001) 27, 241–248. Transplantation of progenitor cells which have been Keywords: PBPC; self-renewal; kinetics; myeloid; mobilised into the bloodstream (PBPC) following the erythroid administration of G-CSF results in more rapid neutrophil recovery than transplantation of bone marrow (BM). The reasons for the accelerated neutrophil engraftment are not clear, but would be explained by Progenitor cells mobilised into the circulation by cytokine increased self-replication of myeloid progenitor cells exposure1,2 or by chemotherapy3–5 are a highly accessible (CFU-GM). We have used a CFU-GM replating assay resource for autologous or allogeneic stem cell transplan- to investigate myeloid progenitor self-replication, and tation. Mobilised peripheral blood progenitor cells (PBPC) quantification of subcolony formation during erythroid differ markedly from progenitors derived from bone mar- burst formation to quantify erythroid progenitor self- row (NBM) by a range of criteria. In particular, 30–60% renewal. Secondary colony formation by CFU-GM, of NBM CD34ϩ cells are in S phase compared with few grown from PBPC and then replated was increased or none in the PBPC CD34ϩ population.6–8 PBPC CD34ϩ compared with secondary colony formation by BM cells are metabolically inactive, as judged by rhodamine CFU-GM (P = 0.0001); erythroid subcolony formation 123 uptake,9,10 and show altered surface antigen and was not altered. There was no difference between the adhesion molecule expression compared with NBM CD34ϩ replating abilities of PBPC CFU-GM derived from allo- cells.11–15 Functional studies indicate that PBPC cells are geneic donors (normal individuals) and autologous enriched for clonogenic cells and for LTC-IC.8,16–19 The donors (patients with malignant disease) although dif- relatively extensive changes induced in PBPC by the mobil- ferences were found between subgroups of autologous isation procedure raised concerns initially that the ability donors. The increased replication of PBPC could not be of these cells to sustain long-term engraftment might be accounted for by a reduction in progenitor cell impaired.9,10 In practise, this has not proved to be the case. apoptosis; PBPC CFU-GM contained slightly fewer Moreover, one of the advantages of PBPC transplants is apoptotic CD34؉ cells than BM CFU-GM. The their ability to significantly accelerate white cell recovery increased replication by PBPC CFU-GM was reversible immediately post transplant compared with bone mar- because it declined when CFU-GM colonies were pass- row.20–24 aged through three sequential CFU-GM replating Very few data are available to explain why PBPC are cycles. This decline in self-replication was more rapid more effective than NBM at initial marrow repopulation. than the decline seen in replated BM CFU-GM. The PBPC recipients are infused with greater numbers of gra- self-replication of PBPC CFU-GM, and subcolony for- nulocyte–macrophage colony-forming cells (CFU-GM) mation by BFU-E could be further enhanced by than NBM recipients. However, when PBPC recipients are exposure to cytokines in vitro. We conclude that mobilis- compared with NBM recipients given the same number of ation alters the replication kinetics of myeloid, but not CFU-GM/kg, neutrophil recovery takes 1.52 Ϯ 0.38 times of erythroid, progenitor cells, that mobilisation-induced as long in the NBM recipients (20.3 Ϯ 1.98 vs 13.8 Ϯ 2.38 events are of limited duration and that in vitro exposure days; P = 0.001).25 This suggests that qualitative rather than quantitative differences between progenitor cells in bone marrow and mobilised blood are responsible for differences in the speed of neutrophil recovery although it should be Correspondence: Dr SB Marley, Leukaemia Research Fund Centre for noted that some authors18–20 have reported a higher pro- Adult Leukaemia, Department of Haematology, Imperial College School portion of primitive haemopoietic progenitors (LTC-IC) in of Medicine, Hammersmith Campus, DuCane Road, London W12 ϩ 25 ONN, UK the PBPC CD34 population. Scott et al found that Received 7 July 2000; accepted 21 October 2000 mobilised progenitor cells were less sensitive than bone PBPC progenitor kinetics SB Marley et al 242 marrow progenitor cells to stromal cell-mediated negative myeloma patients for autologous transplant. Other myel- regulation, which might explain the slower neutrophil oma patients received cyclophosphamide plus 5 ␮g/kg/day recovery in BM recipients. Similarly, PBPC CFU-GM were G-CSF from day 5. Lymphoma patients were mobilised noted to be more responsive than NBM CFU-GM to recom- with etoposide plus 5 ␮g/kg/day G-CSF from day 3 and binant cytokine combinations and to have greater pro- breast cancer patients with FEC (fluorouracil, epirubicin, ductive capacity in vitro.8 However, Prosper et al26,27 found cyclophosphamide) plus 5 ␮g/kg/day G-CSF. that both the frequency and the in vitro productivity of ϩ LTC-IC from NBM and PBPC CD34 cells were compara- CFU-GM assay ble, whilst others showed that the productivities of P⌬ cells (primitive progenitors) and LTCICs from PBPC were Mononuclear cells (MNC) were obtained by density gradi- reduced.28,29 ent centrifugation over Lymphoprep (Nycomed, Oslo, An increase in progenitor cell self-renewal would pro- Norway) and depleted of adherent cells by a 2-h incubation vide an alternative explanation for rapid neutrophil recov- in alpha medium (Gibco, Paisley, UK) plus 15% foetal calf 30–32 ° ery. We have developed in vitro assays to measure the serum in tissue culture plastic flasks at 37 Cin5%CO2 balance between self-renewal and differentiation of haemo- in air. Non-adherent MNC were plated at 1 ϫ 105/ml in poietic progenitor cells. These are based on secondary col- methylcellulose containing serum (Methocult H4230; Met- ony formation by replated CFU-GM colonies for the achem Diagnostics, Northampton, UK) and supplemented myeloid lineage;33 for the erythroid lineage, the assay is with recombinant human cytokines (100 ng/ml G-CSF based on subcolony formation during erythroid burst forma- (Granocyte; Chugai Pharma), 1 ng/ml GM-CSF, 5 ng/ml tion.34 Using these assays, we have demonstrated that self- IL3 and 20 ng/ml SCF (First Link, West Midlands, UK)) renewal in both the erythroid and myeloid progenitor popu- in petri dishes. For some experiments, cytokine support lations varies according to tissue source and is modulated comprised G-CSF alone or in combination with other fac- by external factors such as cytokines.34–36 tors as defined in the text. The cultures were incubated at ° Here, we have used the assays to investigate the self- 37 C in humidified 5% CO2 in air for 7 days, and then replicative capacities of mobilised progenitor cells from colonies consisting of at least 50 cells were scored on an normal individuals and patients with haematological malig- inverted microscope. nancies. We found that self-replication by mobilised CFU- GM, but not by mobilised BFU-E, was temporarily CFU-GM replating assay increased compared to bone marrow. The extent of increased CFU-GM replication varied according to the For each sample investigated, 120 consecutive colonies patients’ disease and self-replication by CFU-GM and consisting of 50 or more cells after 7 days of culture were BFU-E could be (further) increased by exposure to cyto- visualised under a binocular microscope and plucked from kines in vitro. The inference that the superior short-term the methylcellulose using a sterile Eppendorf pipette. Each engraftment of PBPC over BM transplants may be a result colony was vigorously dispersed in a separate well of a 96- of enhanced CFU-GM self-renewal leads to the possibility well microtitre plate containing 100 ␮l of methylcellulose that stimulation of cells with selected cytokine combi- plus serum and G-CSF, GM-CSF, IL3 and SCF as pre- nations in vitro prior to transplantation may increase the viously detailed. The plates were incubated for a further 7 ° speed of neutrophil engraftment and have beneficial effects days at 37 C in humidified 5% CO2 in air. Each well was on red cell reconstitution. scored for the presence and number of secondary colonies of 50 or more cells.

Materials and methods CFU-GM sequential replating Haemopoietic cell samples CFU-GM were grown for 7 days then 120 colonies picked to determine replating capability, as described above. At All samples were obtained with informed consent and local the same time, no fewer than 100 colonies of 50 cells or research ethics committee approval. G-CSF-mobilised per- more were plucked from the methylcellulose using a sterile ipheral blood progenitor cell samples were obtained from Eppendorf pippette and pooled in 3 ml of fresh methylcellu- leukaphereses processed by the Stem Cell Laboratory, lose culture medium. After vigorous dispersal, three 1 ml Hammersmith Hospital, in excess of clinical requirements. aliquots were dispensed into 35 mm2 petri dishes and incu- ° Normal bone marrow was obtained from donations for allo- bated in a humidiﬁed incubator at 37 C with 5% CO2 for geneic transplantation. a further 7 days. After 7 days, 120 individual colonies were again picked to determine AUC and at least 100 colonies PBSC mobilisation protocols were pooled and re-cultured for a further 7 days. The procedure was halted when there were insufﬁcient colonies PBSC were mobilised according to local protocols, using recovered at day 7 to determine the AUC (see below). either glycosylated (Lenograstim; Chugai Pharma, London, UK) or non-glycosylated (Filgrastim; Amgen, Welwyn BFU-E assay Garden City, UK) G-CSF. G-CSF was used alone at 12 ␮g/kg/day for 5 days to mobilise normal donors for allo- Mononuclear cells (MNC) were obtained by density gradi- geneic transplant or at 10 ␮g/kg/day for 5 days to mobilise ent centrifugation over Lymphoprep (Nycomed) and

Bone Marrow Transplantation PBPC progenitor kinetics SB Marley et al 243 depleted of adherent cells by a 2-h incubation in alpha washed with 4 ϫ 500 ␮l cold MM buffer. After removal medium (Gibco) plus 15% foetal calf serum in tissue cul- from the magnet, CD34ϩ cells were eluted from the column ° ture plastic flasks at 37 Cin5%CO2 in air. Non-adherent with 1 ml MM buffer. MNC were plated at 1 ϫ 105 ml in methylcellulose containing serum (Methocult H4230; Metachem Diagnostics) and Determination of apoptosis by TUNEL in CD34ϩ cells supplemented with recombinant human cytokines (3 U/ml from CFU-GM Epo (Eprex; Jenssen-Cilag, High Wycombe, UK), 25 ng/ml IL3 and 100 ng/ml SCF (both First Link)) in petri dishes. CD34ϩ cells suspended in MM buffer were mixed with For some experiments, cytokine support comprised Epo equal volumes of 8% paraformaldehyde for 10 min. The alone or in combination with other factors as defined in the cells were pelleted at 1800 r.p.m. for 5 min then resus- text. The cultures were incubated at 37°C in humidified 5% pended in Dulbecco’s MEM (Gibco) at a concentration of 5 CO2 in air for 14 days. The cultures were then scored for 2 ϫ 10 /ml and 100 ␮l volumes cytospun on to cleaned total BFU-E numbers and the number of sub-colonies microscope slides at 450 r.p.m. for 10 min (Shandon Cyto- determined for at least 120 individual BFU-E. spin 2; Shandon, Pittsburgh, PA, USA). Slides were air- dried overnight, rehydrated in TBS for 15 min at RT and Data analysis for myeloid or erythroid progenitor self- dried. The cells were covered by a 5 ml droplet of protein K replication diluted 1:100 in 10 mm Tris (pH 8), incubated 5 min at RT then dipped three times into TBS and dried. The specimen The data were analysed using a Microsoft Excel spread- was covered with 100 ␮l of the supplied equilibration sheet, version 5.0 (Microsoft, Redmond, WA, USA) or a buffer and incubated for 30 min at RT. Excess buffer was computer program custom-written by Dr NM Blackett. The poured off and freshly prepared TdT labelling mixture raw data are the number (n) of secondary CFU-GM pro- (3 ml TdT enzyme in 57 ml TdT labelling reaction mix duced by each individual primary replated colony or of sub- (Frag-EL; Calbiochem, Nottingham, UK)) was layered on colonies comprising each BFU-E. A cumulative distri- to the cells. The slide was incubated at 37°C in a humidified bution of the proportion of primary CFU-GM producing chamber for 1.5 h then washed ϫ 3 in TBS at RT. A more than n secondary CFU-GM or BFU-E composed of coverslip was applied over mounting medium (Frag-EL) more than n sub-colonies is plotted on a logarithmic y axis. and sealed with nail varnish to prevent evaporation. At least The x axis is expressed as log2n which reflects the number 100 cells from randomly selected fields were scored by of cell doublings required to produce the observed number fluorescent light microscopy (494 nm). Viable cells stained of secondary colonies or sub-colonies. This allows the data blue whilst apoptotic cells appeared as small fragmented to be fitted to a logarithmic curve very precisely. Finally, bodies staining bright green. the area-under-the-curve (AUC) is calculated using the tra- pezium rule. This step is performed because the distribution of secondary colonies per primary CFU-GM and of sub- Statistical analysis colonies per BFU-E is highly skewed so that median, not Statistical analysis was performed using StatView SE ϩ mean, values are appropriate. However, in cases where graphics software for the Macintosh computer (Abacus fewer than 50% of the primary colonies yield secondary Concepts, Berkeley, CA, USA). colonies the median will be zero and not provide any information. The AUC takes account of the proportion of CFU-GM that do not produce any secondary colonies on Results replating. Self-replication by progenitor cells in PBPC collections Separation of CD34ϩ cells from day 7 CFU-GM Figure 1 shows the results obtained in the CFU-GM replat- CFU-GM were grown in methylcellulose for 7 days then ing assay from 56 PBPC harvests and 54 normal bone mar- picked and replated to determine self-replication, as row specimens. Overall, the AUC for the PBPC group was described above. The remaining d7 colonies were pooled markedly and significantly greater than that for the BM into 3 ␮l MM buffer (PBS (Gibco) supplemented with group. In contrast, there was no difference between the 5mm EDTA and 0.5% BSA (Gibco)) and mixed vigorously AUCs derived from PBPC and BM samples in the BFU- to dissolve the methylcellulose and disperse the cells. The E assay (Figure 2). These results show that mobilisation cells were pelleted at 1800 r.p.m. for 5 min. Separation of increases the capacity of myeloid lineage progenitors to CD34ϩ cells was achieved using immunomagnetic beads self-replicate, but has no effect on erythroid progenitor (MiniMacs; Miltenyi Biotec, Camberley, UK). Briefly, the cell kinetics. cell pellet was resuspended in 30 ␮l MM buffer to which The PBPC group consisted of collections from normal was sequentially added 10 ␮l reagent A1 and 10 ␮l reagent donors for allogeneic transplantation and collections from A2. After 15 min at 4°C, the cells were washed with 5 ml patients with malignant disease for autografting. The cold MM buffer and resuspended in 40 ␮l MM buffer. Ten myeloid AUC values for the allogeneic group were 150.3 ␮l reagent B was added, incubated at 4°C for a further Ϯ 20.9 (n = 15) and for the autologous group were 186.3 15 min then the cells washed in 5 ml MM buffer and resus- Ϯ 16.5 (n = 38). These values were not significantly differ- pended in 1 ml MM buffer. The cells were passed down a ent (P = 0.12) (Figure 3). pre-flushed column in a magnetic field which was then Subdivision of the autologous group into collections

Bone Marrow Transplantation PBPC progenitor kinetics SB Marley et al 244 200 allogeneic (normal) and autologous (diseased/treated) PBPC collections is uniformly greater than that of normal marrow CFU-GM. Also, the AUC measured in collections from diseased/treated patients, with the exception of NHL patients, is not lower than that measured for mobilised normal donors, and it may even be higher.

Effects of cryopreservation on CFU-GM self-replication 100 in PBPC harvests Since the majority of PBPC harvests are cryopreserved before infusion, we tested fresh and frozen-thawed samples from five patients. There was a moderate reduction in AUC Ϯ Ϯ Mean AUC of CFU-GM + 1 s.e. n = 54 n = 56 from 193 49 in fresh samples to 101 28 in frozen- thawed samples (P = 0.005; Wilcoxon signed rank test). Thus, cryopreservation appears to reduce significantly the 0 ability of surviving CFU-GM to self-replicate. NBM PBPC (P = 0.0001, MannWhitney U) Increased self-replication by mobilised CFU-GM is not a reflection of reduced apoptosis Figure 1 Self-replication by CFU-GM grown from PBPC or BM. CFU- GM colonies were grown for 7 days then replated individually. Secondary The increased number of secondary CFU-GM in PBPC- colonies were scored after a further 7 days. Self-replication is expressed dervied colonies could result from a reduced loss of pro- as the area-under-the-curve (AUC) of the distribution of secondary colonies per primary replated colony. genitors by apoptosis during primary colony development. The proportion of apoptotic CD34ϩ cells in day 7 CFU- GM was measured by Tunel. This gave the apoptosis level 120 in the progenitor population at the time of replating and so relates directly to the AUC subsequently determined. The ϩ 100 frequency of apoptotic CD34 cells was very low in NBM- derived colonies (2.2 Ϯ 0.3; mean Ϯ s.e; n = 18) and there was only a small, non-significant reduction in apoptosis in 80 PBPC-derived colonies (1.6 Ϯ 0.4%; n = 17; P = 0.13 (Mann–Whitney U)). 60 Increased self-replication by mobilised CFU-GM is ‘self- 40 limiting’ A sustained increase in myeloid progenitor self-replication Mean AUC of BFU-e + 1 s.e.m. 20 could be detrimental and lead to marrow hypercellularity n = 9 n = 15 and raised peripheral blood counts. Accordingly, we inves- 0 tigated the replating ability of successive generations of NBM PBPC PBPC and BM CFU-GM colonies. The data in Figure 4 P = 0.12, MannWhitney U show that the replating abilities of PBPC and BM CFU- GM fall exponentially on sequential replating and that the Figure 2 Self-replication by BFU-E subcolony-forming cells in replating ability of PBPC CFU-GM in fact falls faster than erythroid bursts grown from PBPC or BM. The numbers of subcolonies that of BM CFU-GM. Thus, the self-replication ability of in individual 14-day-old erythroid bursts were scored. Self-replication is CFU-GM declines as a function of the number of cell expressed as the AUC of the distribution of subcolony numbers in individual bursts. generations.

Cytokine exposure increased self-replication by mobilised from patients with breast cancer, myeloma, non-Hodgkin’s myeloid and erythroid progenitors lymphoma and ‘others’ (testicular cancer, germ cell tumour, T cell ALL or Hodgkin’s lymphoma) revealed that the The above results suggest that in vivo exposure to G-CSF myeloid AUC derived from NHL PBPC was significantly increases the self-replicative capacity of CFU-GM in PBPC lower than that from the allogeneic group (P = 0.026) and collections. We have previously shown that IL-3, but not that from the rest of the autologous group (P = 0.005), but SCF, increase the replating ability of BM CFU-GM, and still greater than the AUC measured from BM (P = 0.01). subcolony formation by BFU-E, in vitro.35 To determine The AUC from the ‘others’ was significantly greater than whether PBPC CFU-GM and BFU-E respond in a similar that of the allogeneic group (P = 0.02) and that of the rest fashion we performed the experiments shown in Figures 5 of the autologous group (P = 0.05). and 6. The data show that addition of either IL-3 or SCF to These results show that self-replication by CFU-GM in G-CSF (CFU-GM) or Epo (BFU-E) significantly increased

Bone Marrow Transplantation PBPC progenitor kinetics SB Marley et al 245 300

200

100 Mean AUC of CFU + GM = 1 s.e. 0 NBM All PBPC Allo Auto Breast Myeloma NHL Others n = 54 n = 56 n = 18 n = 38 n = 11 n = 13 n = 8 n = 6 Statistical comparisons (MannWhitney U) Allogeneic vs autologus P = 0.12 NHL vs allogeneic P = 0.026 vs autologous P = 0.005 vs NBM P = 0.01 Other group vs allogeneic P = 0.02 vs autologous P = 0.05

Figure 3 Self-replication by CFU-GM grown from PBPC collections from different groups of PBPC donors. The donors contributing to the data shown in Figure 1 were ﬁrst subdivided according to whether they were allogeneic donors (ie normal) or autologous donors (ie diseased/treated). The autologous PBPC group was then subdivided according to the donor’s disease.

100 ents of PBPC transplants compared with recipients of BM. The speed of neutrophil recovery following PBPC transplantation suggests that neutrophils are derived from pre- existing CFU-GM, rather than from stem cells, in the NBM (n = 6) immediate post-transplant period. However, it may be calculated that 108 CFU-GM, or a total nucleated cell dose in the order of of 1011 would be needed to supply sufficient neutrophils for 1 day.31 Another possibility is that the CFU- 10 GM have a variable capacity to amplify through self- PBPC (n = 10) renewal that is increased when demand for neutrophils is high.31 This idea led us to establish the CFU-GM replating assay in order to measure the self-renewal of CFU-GM in various circumstances.33 The increased replating ability of CFU-GM in PBPC harvests, compared with bone marrow CFU-GM, is therefore consistent with the more rapid rate 1 of neutrophil recovery in recipients of PBPC since whilst AUC relative to the first replate (means + s.e.m.) 01 23a shift in emphasis to self-renewal at progenitor level would Replate number briefly reduce the emergence of neutrophils, the amplification potential between CFU-GM and neutrophil is such Figure 4 Decline in replating ability with successive replating cycles. Primary CFU-GM colonies were grown and replated into 96-well plates that even small numerical changes in the progenitor popu- as before. In addition, batches of primary colonies were replated into petri lation can have a very large effect on neutrophil output. dishes. Secondary colonies in the petri dishes were replated into 96-well Some authorities have proposed that normal steady-state plates (replate number 2). Batches of secondary colonies were also haemopoiesis is accomplished by a considerable loss of replated into petri dishes. These were subsequently replated into 96-well 37 plates (replate number 3). progenitor cells by apoptosis. Accordingly, suppression of apoptotic cell loss by, for example, cytokines could provide a mechanism for progenitor cell expansion or recov- the replating ability of progenitors and that even more marked ery. Our results do not, however, support a reduction in effects are seen when the cytokines are combined. apoptotic cell loss as a mechanism for increasing the AUC measured for PBPC-derived CFU-GM colonies. Another explanation for increased AUC in PBPC might relate to Discussion prior chemotherapy or periods of cytopenia in the autologous donor group, but there was no significant difference This study has revealed qualitative differences between pro- between the total autoPBPC group and the alloPBPC genitor cells in PBPC and BM harvests that may account (normal donor group). Thus, the effects we have observed for the more rapid neutrophil recovery observed in recipi- can be attributed to a large extent to the mobilisation treat-

Bone Marrow Transplantation PBPC progenitor kinetics SB Marley et al 246 P values refer to cytokine G-CSF combinations vs G-CSF alone

G-CSF + IL-3 P = 0.0002

G-CSF + SCF P = 0.025

G-CSF + IL-3 + SCF P = 0.028

G-SCF + SCG + IL-3 + GM-CSF P = 0.005

0 100 200 300 Mean AUC of CFU-GM + 1 s.d.

Figure 5 Effects of cytokines on the replating ability of PBPC CFU-GM. Colonies were grown in G-CSF to provide baseline data because too few colonies for replating are obtained with no cytokine at all. Signiﬁcant increases in AUC were seen when single cytokines or combinations of cytokines were added to G-CSF.

120 population matures with time. The rates of decline are com- parable for mobilised PBPC and NBM. The data therefore 100 suggests that the mobilisation procedure induces a tempor- ary elevation in self-renewal from which the cells then P = 0.0001 decline naturally rather than having a sustained effect 80 which might result in over-production of mature myeloid cells. This is consistent with the data using different cyto- 60 P = 0.03 kines which shows clearly that self-renewal by progenitor cells is very responsive to exogenous control and supports 40 the hypothesis that myeloid homeostasis can be rapidly and reversibly controlled at the progenitor level. The experiments using the erythroid colony assay 20 Mean AUC of BFU-e + 1 s.e.m. showed that the kinetic changes were restricted to the myeloid lineage. This might have been anticipated since 0 mobilisation was accomplished using the myeloid specific Epo Epo + IL-3 Epo + SCF growth factor, G-CSF. However, this result does demon- P values refer to comparison of cytokine combinations strate that the growth factor itself, and not simply the fact against Epo alone (MannWhitney U, n = 16 that progenitor cells circulate, may have an important Figure 6 Effects of IL-3 and stem cell factor (SCF) on sub-colony for- impact on progenitor cell behaviour. The BFU-E population mation during PBPC erythroid burst formation. Erythroid bursts were is mobilised into the blood, but BFU-E kinetics are not grown in Epo to provide baseline data. Significant increases in subcolony altered. In conclusion, we have shown that the self-renewal formation were seen when IL-3 was added to G-CSF, but not when SCF rate of lineage-committed progenitor cells can vary in was added to G-CSF. response to external stimuli such as exogenous cytokines or G-CSF induced mobilisation in vivo. We hypothesise ment rather than to treatment with cytotoxic drugs. PBPC that, in the myeloid series at least, this is a mechanism from NHL patients did have a significantly lower AUC than which allows for rapid increase in the tempo of myelopo- either alloPBPC or the rest of the autoPBPC. This might iesis without placing proliferative stress on the stem cell reflect an effect of etoposide as a mobilisation agent. How- compartment. The transient responsiveness of this mech- ever, the small group of miscellaneous autoPBPC donors anism to exogenous cytokines, plus the demonstrated categorised as ‘others’ were also mobilised with etoposide superiority of PBPC in short-term engraftment, offer the yet had PBPC with an AUC significantly greater than either possibility of artificially enhancing the engraftment capa- alloPBPC or the rest of the autoPBPC group, suggesting bility of progenitor cell infusions prior to transplantation. that the small differences observed are more likely to relate to underlying disease than to mobilisation regime. The AUC of mobilised PBPC, regardless of donor status or Acknowledgements mobilisation regime, remained significantly higher than the AUC of NBM progenitors. The work was supported by the Leukaemia Research Fund of The sequential replating experiments demonstrated that Great Britain. We thank the clinical haematologists and the staff the self-renewal capability of NBM CFU-GM declines of the stem cell laboratory for their assistance in providing the exponentially with serial replating as the progenitor cell samples for this investigation.

Bone Marrow Transplantation PBPC progenitor kinetics SB Marley et al 247 References ating cells) in vitro and their expansion in defined medium. Proc Natl Acad Sci USA 1996; 93: 1470–1474. 1 Welte K, Gabrilove J, Bronchud MH et al. Filgrastim (r- 18 Pettengell R, Luft T, Henschler R et al. Direct comparison by metHug-CSF): the first 10 years. Blood 1996; 88: 1907–1929. limiting dilution analysis of long-term culture-initiating cells 2 Dale DC, Liles WC, Summer WR, Nelson S. Review: gra- in human bone marrow, umbilical cord blood, and blood stem nulocyte colony-stimulating factor – role and relationships in cells. Blood 1994; 84: 3653–3659. infectious diseases. J Infect Dis 1995; 172: 1061–1075. 19 Fujisaki T, Otsuka T, Harada M et al. Granulocyte colony- 3 Kessinger A, Armitage JO. The evolving role of autologous stimulating factor mobilizes primitive hematopoietic stem peripheral stem cell transplantation following high-dose ther- cells in normal individuals. Bone Marrow Transplant 1995; apy for malignancies. Blood 1991; 77: 211–213. 16: 57–62. 4 Korbling M, Holle R, Haas R et al. Autologous blood stem- 20 Gratwohl A, Hermans J, Baldomero H. Blood and marrow cell transplantation in patients with advanced Hodgkin’s dis- transplantation activity in Europe 1995. European Group for ease and prior radiation to the pelvic site. J Clin Oncol 1990; Blood and Marrow Transplantation (EBMT). Bone Marrow 8: 978–985. Transplant 1997; 19: 407–419. 5 Weaver CH, Schwartzberg L, Li W et al. High-dose chemo- 21 Korbling M, Przepiorka D, Huh YO et al. Allogeneic blood therapy and autologous peripheral blood progenitor cell trans- stem cell transplantation for refractory leukemia and lym- plant for the treatment of Hodgkin’s disease. Bone Marrow phoma: potential advantage of blood over marrow allografts. Transplant 1996; 17: 715–721. Blood 1995; 85: 1659–1665. 6 Roberts AW, Metcalf D. Noncycling state of peripheral blood 22 Bensinger WI, Clift R, Martin P et al. Allogeneic peripheral progenitor cells mobilized by granulocyte colony-stimulating blood stem cell transplantation in patients with advanced hem- factor and other cytokines. Blood 1995; 86: 1600–1605. atologic malignancies: a retrospective comparison with mar- 7 Uchida N, He D, Friera AM et al. The unexpected G0/G1 row transplantation. Blood 1996; 88: 2794–2800. cell cycle status of mobilized hematopoietic stem cells from 23 Urbano-Ispizua A, Rozman C, Martinez C et al. Rapid peripheral blood. Blood 1997; 89: 465–472. engraftment without significant graft-versus-host disease after 8 Lemoli RM, Tafuri A, Fortuna A et al. Cycling status of allogeneic transplantation of CD34ϩ selected cells from per- ϩ CD34 cells mobilized into peripheral blood of healthy donors ipheral blood. Blood 1997; 89: 3967–3973. by recombinant human granulocyte colony-stimulating factor. 24 Bacigalupo A, Zikos P, Van Lint MT et al. Allogeneic bone Blood 1997; 89: 1189–1196. marrow or peripheral blood cell transplants in adults with 9 Andrews RG, Singer JW, Bernstein ID. Human hematopoietic ϩ hematologic malignancies: a single-center experience. Exp precursors in long-term culture: single CD34 cells that lack Hematol 1998; 26: 409–414. detectable T cell, B cell, and myeloid cell antigens produce 25 Scott MA, Apperley JF, Bloxham DM et al. Biological proper- multiple colony-forming cells when cultured with marrow ties of peripheral blood progenitor cells mobilized by cyclo- stromal cells. J Exp Med 1990; 172: 355–358. phosphamide and granulocyte colony-stimulating factor. Br J 10 Srour EF, Leemhuis T, Brandt JE et al. Simultaneous use of Haematol 1997; 97: 474–480. rhodamine 123, phycoerythrin, Texas red, and allophycocy- 26 Prosper F, Stroncek D, Verfaillie CM. Phenotypic and func- anin for the isolation of human hematopoietic progenitor cells. tional characterization of long-term culture-initiating cells Cytometry 1991; 12: 179–183. present in peripheral blood progenitor collections of normal 11 Mohle R, Murea S, Kirsch M, Haas R. Differential expression donors treated with granulocyte colony-stimulating factor. of L-selectin, VLA-4, and LFA-1 on CD34ϩ progenitor cells Blood 1996; 88: 2033–2042. from bone marrow and peripheral blood during G-CSF- 27 Prosper F, Vanoverbeke K, Stroncek D, Verfaillie CM. Primi- enhanced recovery. Exp Hematol 1995; 23: 1535–1542. tive long-term culture initiating cells (LTC-ICs) in granulocyte 12 Masauzi N, Kobayashi M, Suzuki G et al. Comparative studies colony-stimulating factor mobilized peripheral blood progeni- of differentiation antigens and adhesion molecules expressed ϩ ϩ tor cells have similar potential for ex vivo expansion as primi- on CD34 cells staying in bone marrow with those on CD34 cells in peripheral blood mobilized by two different protocols tive LTC-ICs in steady state bone marrow. Blood 1997; 89: with chemotherapy/rhG-CSF or steady state rhG-CSF. Blood 3991–3997. 1995; 86 (Suppl. 1): Abstr. 413. 28 Scott MA, Apperley JF, Jestice HK et al. Plastic-adherent pro- 13 Mohle R, Haas R, Hunstein W. Expression of adhesion mol- genitor cells in mobilised peripheral blood progenitor cell col- ecules and c-kit on CD34ϩ hematopoietic progenitor cells: lections. Blood 1995; 86: 4468–4473. comparison of cytokine-mobilized blood stem cells with nor- 29 Sutherland HJ, Eaves CJ, Lansdorp PM, Hogge DE. Kinetics mal bone marrow and peripheral blood. J Hematother 1993; of committed and primitive blood progenitor cell mobilisation 2: 483–489. after chemotherapy and growth factor treatment and their use 14 To LB, Haylock DN, Dowse T et al. A comparative study of in autotransplants. Blood 1994; 83: 3808–3814. the phenotype and proliferative capacity of peripheral blood 30 Gordon MY, Blackett NM. Some factors determining the (PB) CD34ϩ cells mobilized by four different protocols and minimum number of cells required for successful clinical those of steady-phase PB and bone marrow CD34ϩ cells. engraftment. Bone Marrow Transplant 1995; 15: 659–662. Blood 1994; 84: 2930–2939. 31 Gordon MY, Blackett NM. Reconstruction of the hematopo- 15 Bellucci R, de Propris MS, Buccisano F et al. Modulation of ietic system after stem cell transplantation. Cell Transplan- VLA-4 and L-selectin expression on normal CD34ϩ cells dur- tation 1998; 7: 339–344. ing mobilization with G-CSF. Bone Marrow Transplant 1999; 32 Gordon MY, Blackett NM, Lewis JL, Goldman JM. Evidence 23: 1–8. for a mechanism that can provide both short-term and long- 16 Tjonnfjord GE, Steen R, Evensen SA et al. Characterization term haemopoietic repopulation by a seemingly uniform popu- of CD34ϩ peripheral blood cells from healthy adults mobil- lation of human haemopoietic precursor cells. Leukemia 1995; ized by recombinant human granulocyte colony-stimulating 9: 1252–1256. factor. Blood 1994; 84: 2795–2801. 33 Gordon MY, Marley SB, Lewis JL et al. Treatment with inter- 17 Petzer AL, Hogge DE, Landsdorp PM et al. Self-renewal of feron-␣ preferentially reduces the capacity for amplification primitive human hematopoietic cells (long-term-culture-initi- of granulocyte–macrophage progenitors (CFU-GM) from

Bone Marrow Transplantation PBPC progenitor kinetics SB Marley et al 248 patients with chronic myeloid leukemia but spares normal 35 Lewis JL, Marley SB, Blackett NM et al. Interleukin-3 (IL-3) CFU-GM. J Clin Invest 1998; 102: 710–715. but not stem cell factor (SCF) increases self-renewal by human 34 Marley SB, Lewis JL, Goldman JM, Gordon MY. Abnormal erythroid burst-forming units. Cytokine 1998; 10: 49–54. kinetics of colony formation by erythroid burst-forming units 36 Lewis JL. PhD thesis 1998, University of London. (BFU-E) in chronic myeloid leukaemia. Br J Haematol 1996; 37 Koury MJ. Programmed cell death (apoptosis) in hematopo- 93: 878–883. iesis. Exp Hematol 1992; 20: 391–394.

Bone Marrow Transplantation